Digital fingerprints generated from coil brazing

Abstract

A system is configured to receive video footage of evaporator coil slabs after they exit an automated coil brazer. The system is further configured to convert the video footage to greyscale and isolate frames from the greyscale video footage. Each frame comprises an image of a different evaporator coil slab. The system is further configured to generate a first digital fingerprint comprising a binary feature vector for each point in a first subset of feature points from the first frame, and generate a second digital fingerprint comprising a binary feature vector for each point in a second subset of feature points from the second frame.

The invention claimed is: 
     
         1 . A system for generating digital fingerprints from coil brazing, comprising:
 a first camera configured to record a first video footage of a plurality of evaporator coil slabs after the plurality of evaporator coils exit an automated coil brazer;   an analysis tool comprising a hardware processor communicatively configured to:
 receive the first video footage from the first camera, the first video footage comprising the plurality of evaporator coil slabs; 
 convert the first video footage to a first greyscale video footage; 
 isolate, from the first greyscale video footage a plurality of frames, comprising:
 a first frame comprising a first image of a first evaporator coil slab, the first evaporator coil slab comprising a first plurality of brazed tube junctions; 
 a second frame comprising a second image of a second evaporator coil slab, the second evaporator coil slab comprising a second plurality of brazed tube junctions; 
 
 generate a first digital fingerprint comprising a binary feature vector for each point in a first subset of feature points from the first frame; and 
 generate a second digital fingerprint comprising a binary feature vector for each point in a second subset of feature points from the second frame. 
   
     
     
         2 . The system of  claim 1 , further comprising:
 a second camera coupled to the hardware processor and configured to record a second video footage of the plurality of evaporator coils in a first assembled state, wherein the first assembled state comprises two evaporator coil slabs joined together in a “V” shape; and   wherein the hardware processor is further configured to:
 receive the second video footage from the second camera; 
 convert the second video footage to a second greyscale video footage; 
   isolate, from the second greyscale video footage, a third frame comprising a third image of an evaporator coil in the first assembled state;   determine that a first coil slab in the evaporator coil in the first assembled state is associated with the first digital fingerprint and that a second coil slab in the evaporator coil is associated with the second digital fingerprint; and   generate a third digital fingerprint comprising the first and second digital fingerprints.   
     
     
         3 . The system of  claim 2 , further comprising:
 a third camera coupled to the hardware processor and configured to record a third video footage of evaporator coils in a second assembled state, wherein the second assembled state comprises the first assembled state and additional metal tubing on an exterior of one or both of the evaporator coil slabs; and   wherein the hardware processor is further configured to:
 receive the third video footage from the third camera; 
 convert the third video footage to a third greyscale video footage; 
   isolate, from the third greyscale video footage, a fourth frame comprising a fourth image of the evaporator coil in the second assembled state;   identify a plurality of feature points in the fourth frame, the feature points comprising corners;   determine that a subset of feature points selected from the plurality of feature points identified in the fourth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generate a fourth digital fingerprint comprising a binary feature vector for each point in a subset of feature points selected from a plurality of feature points identified in the fourth frame;   determine that the evaporator coil in the second assembled state is associated with the third digital fingerprint; and   update a database to link the third digital fingerprint with the fourth digital fingerprint.   
     
     
         4 . The system of  claim 3 , further comprising:
 a fourth camera coupled to the hardware processor and configured to record a fourth video footage of evaporator coils in a third assembled state, wherein the third assembled state comprises the second assembled state wherein the additional metal tubing has been brazed; and   wherein the hardware processor is further configured to:
 receive the fourth video footage from the fourth camera; 
 convert the fourth video footage to a fourth greyscale video footage; 
   isolate, from the fourth greyscale video footage, a fifth frame comprising a fifth image of the evaporator coil in the third assembled state;   identify a plurality of feature points in the fifth frame, the plurality of feature points comprising corners;   determine that a subset of feature points selected from the plurality of feature points identified in the fifth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generate a fifth digital fingerprint comprising a binary feature vector for each point in a subset of feature points selected from a plurality of feature points identified in the fifth frame;   determine that the evaporator coil in the third assembled state is associated with one of the third digital fingerprint or the fourth digital fingerprint; and   update the database to link the fifth digital fingerprint with the second digital fingerprint, the third digital fingerprint, and the fourth digital fingerprint.   
     
     
         5 . The system of  claim 4 , wherein:
 the second camera is located further along a production line than the first camera;   the third camera is located further along the production line than the second camera; and   the fourth camera is located further along the production line than the third camera.   
     
     
         6 . A method for generating digital fingerprints from coil brazing, comprising:
 receiving a first video footage from a first camera configured to record video of a plurality of evaporator coil slabs after the plurality of evaporator coil slabs exit an automated coil brazer, the footage comprising the plurality of evaporator coil slabs;   converting the first video footage to a first greyscale video footage;   isolating, from the first greyscale video footage a plurality of frames, comprising:
 a first frame comprising a first image of a first evaporator coil slab, the coil slab comprising a plurality of brazed tube junctions; 
 a second frame comprising a second image of a second evaporator coil slab, the coil slab comprising a plurality of brazed tube junctions; 
   generating a first digital fingerprint comprising a binary feature vector for each point in a first subset of feature points from the first frame; and   generating a second digital fingerprint comprising a binary feature vector for each point in a second subset of feature points from the second frame.   
     
     
         7 . The method of  claim 6 , further comprising:
 receiving a second video footage from a second camera configured to record video of evaporator coils in a first assembled state, wherein the first assembled state comprise two evaporator coil slabs joined together in a “V” shape;   converting the second video footage to a second greyscale video footage;   isolating, from the second greyscale video footage, a third frame comprising an image of an evaporator coil in the first assembled state;   determining that a first coil slab in the evaporator coil in the first assembled state is associated with the first digital fingerprint and that a second coil slab in the evaporator coil is associated with the second digital fingerprint; and   generating a third digital fingerprint comprising the first digital fingerprint and the second digital fingerprint.   
     
     
         8 . The method of  claim 7 , further comprising:
 receiving a third video footage from a third camera configured to record video of evaporator coils in a second assembled state, wherein the second assembled state comprises the first assembled state and additional metal tubing on an exterior of one or both of the evaporator coil slabs;   converting the third video footage to a third greyscale video footage;   isolating, from the third greyscale video footage from the third camera, a fourth frame comprising an image of the evaporator coil in the second assembled state;   identifying a plurality of feature points in the fourth frame, the feature points comprising corners;   determining that a subset of feature points selected from the plurality of feature points identified in the fourth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generating a fourth digital fingerprint comprising a binary feature vector for each point in a subset of feature points selected from a plurality of feature points identified in the fourth frame;   determining that the evaporator coil in the second assembled state is associated with the third digital fingerprint; and   updating a database to link the third digital fingerprint with the fourth digital fingerprint.   
     
     
         9 . The method of  claim 8 , further comprising:
 receiving a fourth video footage from a fourth camera configured to record video of evaporator coils in a third assembled state, wherein the third assembled state comprises the second assembled state wherein the additional metal tubing has been brazed;   converting the fourth video footage to a fourth greyscale video footage;   isolating, from the fourth greyscale video footage, a fifth frame comprising an image of the evaporator coil in the third assembled state;   identifying a plurality of feature points in the fifth frame, the feature points comprising corners;   determining that a subset of feature points selected from the plurality of feature points identified in the fifth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generating a fifth digital fingerprint comprising a binary feature vector for each point in a subset of feature points selected from a plurality of feature points identified in the fifth frame;   determining that the evaporator coil in the third assembled state is associated with one of the third digital fingerprint or the fourth digital fingerprint; and   updating the database to link the fifth digital fingerprint with the second digital fingerprint, the third digital fingerprint, and the fourth digital fingerprint.   
     
     
         10 . A non-transitory computer readable medium storing instructions that when executed by a hardware processor cause the hardware processor to:
 convert a first video footage, received from a first camera configured to record video of a plurality of evaporator coil slabs after the plurality of evaporator coil slabs exit an automated coil brazer, to a first greyscale video footage;   isolate, from the first greyscale video footage:
 a first frame comprising a first image of a first evaporator coil slab, the coil slab comprising a plurality of brazed tube junctions; 
 a second frame comprising a second image of a second evaporator coil slab, the coil slab comprising a plurality of brazed tube junctions; 
   generate a first digital fingerprint comprising a binary feature vector for each point in a first subset of feature points from the first frame; and   generate a second digital fingerprint comprising a binary feature vector for each point in a second subset of feature points from the second frame.   
     
     
         11 . The non-transitory computer readable medium of  claim 10 , wherein the instructions when executed by the hardware processor cause the hardware processor to:
 convert a second video footage, received from a second camera configured to record video of the plurality of evaporator coils in a first assembled state, wherein the first assembled state comprise two evaporator coil slabs joined together in a “V” shape, to a second greyscale video footage;   isolate, from the second greyscale video footage, a third frame comprising a third image of the plurality of evaporator coils in the first assembled state;   determine that a first coil slab in the plurality of evaporator coils in the first assembled state is associated with the first digital fingerprint and that a second coil slab in the plurality of evaporator coils is associated with the second digital fingerprint; and   generate a third digital fingerprint comprising the first digital fingerprint and the second digital fingerprint.   
     
     
         12 . The non-transitory computer readable medium of  claim 11 , wherein the instructions when executed by the hardware processor cause the hardware processor to:
 convert a third video footage, received from a third camera configured to record video of the plurality of evaporator coils in a second assembled state, wherein the second assembled state comprises the first assembled state and additional metal tubing on an exterior of one or both of the plurality of evaporator coil slabs, to a third greyscale video footage;   isolate, from the third greyscale video footage, a fourth frame comprising a fourth image of the evaporator coil in the second assembled state;   identify a plurality of feature points in the fourth frame, the feature points comprising corners;   determine that a subset of feature points selected from the plurality of feature points identified in the fourth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generate a fourth digital fingerprint comprising a binary feature vector for each point in the subset of feature points selected from the plurality of feature points identified in the fourth frame;   determine that the plurality of evaporator coils in the second assembled state is associated with the third digital fingerprint; and   update a database to link the third digital fingerprint with the fourth digital fingerprint.   
     
     
         13 . The non-transitory computer readable medium of  claim 12 , wherein the instructions when executed by the hardware processor cause the hardware processor to:
 convert a fourth video footage, received from a fourth camera configured to record video of the plurality of evaporator coils in a third assembled state, wherein the third assembled state comprises the second assembled state wherein the additional metal tubing has been brazed, to a fourth greyscale video footage;   isolate, from the fourth greyscale video footage, a fifth frame comprising a fifth image of the plurality of evaporator coils in the third assembled state;   identify a fifth plurality of feature points in the fifth frame, the fifth plurality of feature points comprising corners;   determine that a subset of feature points selected from the plurality of feature points identified in the fifth frame are rotationally invariant, wherein a point is rotationally invariant if the point remains identifiable in images of the coil taken from a different angle;   generate a fifth digital fingerprint comprising a binary feature vector for each point in the subset of feature points selected from the plurality of feature points identified in the fifth frame;   determine that the plurality of evaporator coils in the third assembled state is associated with one of the third digital fingerprint or the fourth digital fingerprint; and   update a database to link the fifth digital fingerprint with the second digital fingerprint, the third digital fingerprint, and the fourth digital fingerprint.